Attachment and post-traumatic stress disorder in french military personnel: characterization and exploration of biopsychosocial factors — the at-home study protocol

Abstract

Introduction

In military service members, post-traumatic stress disorder (PTSD) often becomes chronic, with symptoms that significantly interfere with daily life and social functioning. These difficulties pose a major challenge for the rehabilitation and professional reintegration of injured service members. Our protocol describes a study that investigates emotional contagion processes in military personnel with PTSD, in relation to attachment style, along with selected biophysiological and psychosocial factors.

Method

The method compares two groups of French military personnel: one group with PTSD (n = 41) and a control group without PTSD (n = 53). During a single visit, each participant completes four steps: (1) provision of informed consent; (2) saliva collection (for the assay of oxytocin and vasopressin, and for genetic polymorphism analyses); (3) a computer-based emotion recognition task combined with physiological recordings of electrodermal activity and heart rate variability; and (4) completion of a set of validated self-report questionnaires measuring attachment, emotional contagion, empathy, post-traumatic symptoms, stress, adverse childhood experiences, and positive mental health; and a set of sociodemographic questions gathering information on participants’ personal and professional background.

Ethics

Ethical approval has been obtained from the Sud-Est VI Ethics Committee (reference no 25.0555.000354, 14/03/2025).

Trial registration number

ClinicalTrials.gov identifier: NCT06996275.

Graphical abstract

Specifications table

Subject area	Psychology
More specific subject area	Cognitive psychology – Clinical psychology - Neuropsychology
Name of your protocol	AT-HOME: ATtacHment and pOst-traumatic stress disorder in Military personnel: characterization of the french military population, Exploration of biopsychosocial factors, and investigation of the impact of security priming on emotional contagion abilities
Reagents/tools	Biological materials: Sarstedt Salivette devices (reference 51.1534) and ORAGENE saliva collection kit (DNA Genotek, OG-500). Computer task: standard personal computer with 27-inch LED monitor, task developed in-house in C#. Physiological recording: BioPac MP160 system (BioPac Systems, Inc., CEROM, Paris, France). Data acquired using AcqKnowledge 5.0 at a sampling frequency of 1000 Hz. Psychometric Questionnaires: RQ (Relation Questionnaire - attachment), ECS (Emotional Contagion Scale), IRI (Interpersonal Reactivity Index - empathy), ITQ (International Trauma Questionnaire - PTSD and C-PTSD), PSS (Perceived Stress Scale), ACE-Q (Adverse Childhood Experiences Questionnaire), and MHC-SF (Mental Health Continuum Short Form - positive mental health).
Experimental design	Each participant completes a single 90-minute session structured in four steps: −Consent: Each volunteer is received by the principal investigator, who explains the legal framework, what is involved for the participant, and answers any questions. If the person agrees to participate, they are asked to sign consent forms for both the main study and the genetic study. −Biological Samples: Two biological samples are collected: one for salivary oxytocin and vasopressin assays, and the other for genetic polymorphism analyses. The first sample is collected using Salivette; the participant is asked to keep a cotton swab in their mouth. The second uses saliva expectoration. −Experimental Task: The participant completes a computer-based task consisting of two blocks of 180 trials. Each block is preceded by a neutral or secure 5-minute priming period. Photographs of facial expressions (joy, anger, or neutral) are presented for 200 milliseconds, and the participant is asked to identify, as quickly as possible, the emotion being displayed. Physiological measures of the autonomic nervous system are recorded via three electrodes placed on the participant’s chest, and two finger electrodes on their non-dominant hand. −Questionnaires: Finally, the participant completes a series of psychometric questionnaires and a series of sociodemographic questions covering military life and personal information.
Trial registration	Clinical trial: NCT06996275
Ethics	Favorable ethical approval was granted by the Southeast VI Personal Protection Committee on March 14, 2025. All participants received both collective and individual information. Written informed consent is obtained, and forms are stored separately from the study data.
Value of the Protocol	(1) Provides a detailed characterization of social functioning in French military personnel, both with and without PTSD/C-PTSD. (2) Combines psychometric, physiological, and genetic measures to offer a comprehensive understanding of mechanisms underlying social functioning in PTSD. (3) Investigates how secure priming influences the detection and processing of emotional stimuli.

Background

Post-traumatic stress disorder (PTSD) is a psychiatric disorder that can develop following exposure to a traumatic event (e.g., death, serious injury, accident, war, violence…). According to DSM-5 [1], PTSD symptoms are grouped into four clusters: intrusion symptoms, avoidance, negative alterations in cognition and mood, and marked alterations in arousal and reactivity. Its prevalence in the general population is estimated to be around 9% [1], but some authors report rates up to 30% in high-risk populations [17,20,27].

Recently, the ICD-11 introduced complex PTSD [30], characterized by PTSD symptoms and additional disturbances in self-organization, including emotional dysregulation, negative self-concept, and difficulty maintaining relationships. One of the main factors contributing to the development of C-PTSD is repeated or prolonged exposure to traumatic events, particularly during childhood. Adverse childhood experiences, including abuse, neglect, or household dysfunction, are known risk factors for developing C-PTSD later in life [18].

Military personnel are frequently exposed to operational constraints that place significant strain on the body. As a result, their risk of developing stress-related psychopathologies such as PTSD and C-PTSD may be higher than in the general population. Additionally, cultural stigma and operational demands may delay help-seeking, and some symptoms, such as nightmares, can be considered “normal”. Their symptoms often become chronic [28], posing a significant challenge for military healthcare systems. Moreover, access to mental health care varies by country. In France, for example, recognized service-related injuries may qualify individuals for support through the military’s Wounded Support Unit (Cellule d’Aide aux Blessés; CABAT). One of the objectives of the CABAT is to promote social and occupational reintegration, in addition to providing psychiatric care.

Psychologically injured service members frequently report that it is difficult to resume everyday activities such as going to the supermarket, bakery, or cinema, attending an event at a stadium, or even using public transportation. The common feature of these situations is immersion in a social environment. This raises questions about the functioning of social cognition in military personnel with PTSD and C-PTSD. Key components such as attachment processes and core emotional skills, which contribute to social cognition, are often found to be impaired in PTSD.

In this context, it is important to consider the functioning of neuropeptides such as oxytocin and vasopressin (levels and polymorphisms). These two antagonistic neuropeptides are involved in social behavior [8,29], and show altered functioning in PTSD [7,13,16]. Surprisingly, emotional contagion, one of the first steps in perceiving and sharing social information with others [15] has not yet been thoroughly investigated in PTSD. This is an important issue for two reasons. First, the question concerns the physiobiological response of emotional contagion and its link with attachment, and second, little is known about the impact of the safety of the social context on emotional contagion. A better understanding of emotional contagion in relation to attachment and biopsychosocial functioning, depending on the safety of the context, may help to improve social functioning in PTSD, especially military personnel with chronic PTSD.

Description of protocol

Objectives: The main objective of this protocol is to characterize attachment styles in the French military population by comparing a group of military personnel without PTSD (No-PTSD group) to a group of military personnel with PTSD (PTSD group), including those with complex PTSD (C-PTSD).

The secondary objectives are:

• (2.1.) To characterize the French military population with and without PTSD, including C-PTSD, from a biopsychological and a life history perspective.

• (2.2.) To assess the extent to which attachment style moderates’ performance in an emotion recognition task that evaluates emotional contagion processes.

• (2.3.) To investigate associations between oxytocin and vasopressin (both salivary concentrations and genetic polymorphisms) and key psychosocial variables.

• (2.4.) To evaluate autonomic nervous system functioning measured as heart rate variability (HRV) and electrodermal activity (EDA) under secure versus neutral priming conditions, and during recovery after the experimental task.

• (2.5.) To assess the impact of priming on the speed and accuracy of emotional facial expression recognition.

Hypothesis: We hypothesize that military personnel with PTSD will show psychosocial profiles indicative of reduced social functioning: a higher prevalence of insecure attachment style, altered empathic abilities, and lower oxytocin levels associated with higher vasopressin levels. We expect this alteration to be more pronounced among participants presenting symptoms of disturbances in self-organization, as observed in C-PTSD.

Furthermore, we hypothesize that secure priming will improve both the speed and accuracy of emotional facial expression recognition during the experimental task, compared to the neutral priming condition.

Participants and recruitment: The protocol compares two groups of participants: one diagnosed with PTSD (PTSD group) and one without PTSD (No-PTSD group). Sample size calculations are used to determine the required number of participants. As no data are available in the literature regarding the prevalence of attachment styles in the French military population, data are drawn from studies conducted by our research team (in press), which found the following rates: 17% secure attachment in military personnel with PTSD, and 51% in those without PTSD ([25,11].

Given the following parameters:

• •Statistical test: two-tailed
• •Prevalence of secure attachment in the No-PTSD group: 51%
• •Prevalence of secure attachment in the PTSD group: 17%
• •Statistical significance = 0.05
• •Power = 0.90

36 participants per group are needed.

As the study involves a single visit, attrition is not expected. However, it should be noted that some military personnel recruited into the No-PTSD group may suffer from undiagnosed PTSD (despite their operational status). Therefore, 12 additional participants are recruited (corresponding to the 30% prevalence reported in the background). Participants whose score exceeds the PTSD threshold set in the ITQ questionnaire are excluded from the subsequent statistical analyses. This results in 36 participants in the PTSD group and 48 in the No-PTSD group.

Furthermore, given the low proportion of women in the military, and the well-documented differences in social cognitive functioning between men and women reported in the literature, data from female participants are analyzed only if each group includes at least five women. Including an estimated five female participants per group increases the total sample size from 84 to 94 participants.

In our study, the No-PTSD group is recruited from the 92nd Infantry Regiment of the French Army, based in Clermont-Ferrand (France). The PTSD group is recruited during rehabilitation programs for injured service members, organized by the CABAT. Inclusion and exclusion criteria for each group are presented in Table 1.

Table 1.

Inclusion and exclusion criteria.

Inclusion criteria common to both groups
Aged 18–65
Male or female
Provision of written informed consent
Affiliated with the French social security system
Fluent in French

Inclusion criteria specific to the PTSD group
Active-duty military personnel formally diagnosed with post-traumatic stress disorder (PTSD) in the context of their professional service.
Participation in programs organized by the CABAT

Exclusion criteria common to both groups
Adults under legal guardianship
Individuals who decline to participate in the study
French Foreign Legion member with unverified administrative status

Study design

Before the start of the study, all participants—regardless of group—attend a collective information session. They are informed about the context of the study, the regulatory framework in which it is conducted, the nature of their participation, and their rights.

For each participant, the study is divided into four steps, as shown in Fig. 1. Whenever possible, these steps are administered in the same order for all participants. However, operational constraints may make it necessary to adjust the order of steps 2, 3 and 4.

• 1. Individual information and provision of written informed consent

Fig. 1.

Graphical representation of the four steps participants are asked to complete.

In this step, the participant meets the principal investigator, who explains the study procedures in detail and answers any questions. The investigator also verifies all inclusion and exclusion criteria. The participant is asked to sign two informed consent forms: one for the main study and one for the genetic study. The person may decline the genetic analysis while still taking part in the rest of the study. Certain medical data are also collected at this stage, including the year of PTSD diagnosis and the year of the event identified by the participant as the most traumatic. Any history of traumatic brain injury and current medication are also recorded.

• 2. Biological sample collection and processing

After informed consent is obtained, two biological samples are collected, one for the main study and one for the genetic study. Participants are instructed not to smoke or vape, drink, or eat for at least 30 min prior to sample collection. The first is a 2 ml saliva sample collected using a Sarstedt Salivette device (reference 51.1534), intended for the assay of salivary oxytocin and vasopressin. The participant is asked to keep 2 cotton swabs in their mouth for at least 2 min, without chewing excessively to minimize stimulation artifacts. A second 2 ml sample is collected using an ORAGENE device (DNA Genotek, OG-500) for the genetic analysis of oxytocin and vasopressin polymorphisms. The targeted polymorphisms are rs53576 and rs2254298 for the oxytocin receptor gene, and AVPR1A RS1 and AVPR1B for the vasopressin system.

Following collection, the cotton swabs from the first sample are processed using a centrifuge (3200 rpm, 3 min). The extracted material is pipetted into Eppendorf tubes and stored in a cooler at −20 °C until the return from the field. The ORAGENE kits are also stored at −20 °C. Upon return, all samples are transferred to a − 80 °C freezer for long-term storage prior to analysis. Data processing was carried out in accordance with the manufacturer’s protocols for Oxytocin (ADI-901–153A, Enzo) and Vasopressin (ADI-900–017A, Enzo) kits.

For each participant, the time of collection, volume of saliva obtained and current medication are recorded. The time of the samples collection (morning vs afternoon) is being taken into analysis.

• 3. Computer-based task with physiological recording

Description of the task: The task involves the identification of emotional facial expressions. First, each participant completes a 30-trial training phase, followed by 360 target trials divided into two blocks of 180 trials each. A trial consists of a fixation cross presented for 500 ms, followed by the target image for 200 ms, and then a black screen until the participant responds or for a maximum of 1500 milliseconds. Participants are asked to identify the displayed emotion (joy, anger, or neutral) using a keyboard, responding as quickly as possible while minimizing errors. A 5-minute priming phase precedes each block.

HRV and EDA are continuously recorded throughout the task. Additionally, there is a 5-minute recording period both prior to the training phase (baseline in Fig. 2) and after the final trials (recovery in Fig. 2).

Fig. 2.

Workflow of the computer-based task.

Priming: Before each trial block, the participant is exposed to either a neutral or secure priming condition. Neutral priming consists of simple cognitive exercises, such as general knowledge questions or basic problem-solving tasks (e.g., calculations or Tower of London-style tasks). Prior to the AT-HOME study, these tasks were pre-tested in a population similar to the target sample to ensure they do not present any difficulty. The neutral priming is introduced by an audio recording using the same voice as the secure priming. Secure priming consists of a 5-minute audio recording, produced by a professional. The recording guides participants mentally visualize a situation (place, people, or event - determined thanks to a precedent pre-test in the same population conducted by our team) that evokes a sense of safety. Both priming conditions last five minutes and are strictly standardized between participants. After the computer-based task, each participant completes a Visual Analog Scale (VAS) to rate his/her feeling of safety during the secure priming.

The order of priming conditions is counterbalanced across participants.

Images and emotions: In the experimental task, participants are presented with photographs of emotional facial expressions drawn from the Chicago Face Database [22]. Selected images depict neutrality, closed-mouth joy, or anger. A total of 120 images per emotion are used, with each actor displaying all three emotional expressions. The 360 photos are presented to the participant in two blocks of 180 trials. Within each block, the participant views 60 images of each emotion, with randomized presentation. Overall, each participant views all 360 photos, divided into two equal blocks, with each emotion presented an equal number of times. The images used during the training phase do not appear in the target trials.

Application and materials: The application used to run the task was developed by the Département Innovation Numérique et Intelligence Artificielle, a support team from the Armed Forces Biomedical Research Institute (Institut de Recherche Biomédicale des Armées). The task is executed on a standard personal computer with sufficient processing power to run the application smoothly. Visual stimuli are displayed on a 27-inch LED monitor (2560 × 1440 pixels, IPS panel, 1 ms response time, 165 Hz refresh rate). The participant interacts with the application using a modified numeric keypad equipped with four keys: one for each emotion, and a “start” button to resume the block after each break (Fig. 3).

Fig. 3.

Overview of what the participant sees during the task.

Physiological response assessment: Physiological data are recorded using the BioPac MP160 system (BioPac Systems Inc., CEROM, Paris, France). Data are acquired using AcqKnowledge 5.0 software at a sampling frequency of 2000 Hz. EDA is recorded using two electrodes placed on the fingers of the non-dominant hand, while ECG activity is recorded using three electrodes placed on the chest. To minimize signal interference, participants are instructed to remain as still as possible throughout the task.The physiological analysis is performed using the publicly available neurophysiological toolkit, NeuroKit2 [23]. Concerning HRV treatment, a Butterworth filter with a lowcut at 0,5 Hz is applied to remove baseline wander caused by the participant breathing or movement. Then the powerline noise is removed using a moving average filter. If missing R peaks, an interpolation is applied, based on the length of the window (to interpolate) divided by the mean of the effective RR interval. It results in an integer part that represents the number of R peaks to interpolate. This correction is validated by a visual inspection. Heart rate (HR), RMSSD (root mean square of successive differences), SDNN (standard deviation of NN intervals), low-frequency power (LF), high-frequency power (HF), and the LF/HF ratio are computed to assess heart rate variability (HRV). For EDA analysis, the signal is downsampled to 300 Hz and a Butterworth filter is applied with a highcut at 1 Hz. EDA was analyzed by the decomposition of the signal into phasic and tonic component using the gold standard cvxEDA method described by Greco et al. [14] implemented in NeuroKit2. The EDA detection follows the approach by Nabian et al. [24]. Detected EDAs with amplitudes <10% of the maximum EDA amplitude are eliminated.

• 4. Self-report questionnaires

In the final step of the procedure, the participant completes a set of sociodemographic questions and a battery of validated psychometric questionnaires. These are administered on a tablet using the OMR Manager application. Sociodemographic questions focus on both personal and military background: age, sex, highest educational qualification, marital status, living arrangements (geographical separation, on-base housing, private residence, other), number of children, branch of service (Air Force, Navy, Gendarmerie, Army, or Joint Services), length of service, rank, and number of military deployments.

Next, the following are assessed:

Psychosocial assessment

• −Attachment style (RQ, [2]): The RQ measures attachment both categorically and dimensionally. In the first part, participants choose the statement that best describes them from four options corresponding to the four attachment styles. In the second part, they rate each of these statements on a 7-point Likert scale.
• −Emotional contagion (ECS, [9]): The ECS is a self-report measure of emotional contagion. It consists of 15 items that are rated on a 5-point Likert scale. A global score can be calculated, along with five subscales: love and happiness (positive emotions), and fear, sadness, and anger (negative emotions). A French validation is currently in progress within our research team.
• −Empathy (IRI, [6]): The IRI is the gold standard self-report measure of empathy. We use the shortened 15-item version validated in French by Braun et al. [3]. Participants respond on a 5-point Likert scale. Four subscales can be calculated: fantasy and perspective taking (cognitive empathy), and empathic concern and personal distress (affective empathy).

Traumatology assessment

• −Traumatic symptomatology (including C-PTSD) (ITQ, [4,26]): The ITQ aims to assess the severity of post-traumatic and complex post-traumatic symptomatology. First, participants indicate the event they are referring to and when it occurred. Then, they respond to 18 items related to PTSD symptoms, using a 5-point Likert scale.
• −Adverse childhood experiences (ACE-Q, [12]): The ACE-Q measures various adverse life events that may have been experienced during childhood. Participants respond yes or no, based on their personal experience.

Assessment of the participant’s state at the time of the session

• −Perceived stress (PSS-14, [5,21]): The PSS is a self-report questionnaire designed to measure perceived stress over the past month. It uses a 5-point Likert scale. We consider both the global score as well as the two subscales: perceived self-efficacy and perceived helplessness.
• −Positive mental health (MHC-SF, [19],[10]): The MHC-SF aims to assess overall psychological well-being across three subscales: psychological well-being, emotional well-being, and social well-being. It consists of 14 items rated on a 6-point Likert scale.

Statistics

Missing data: Because the study involves a single visit, very little missing data is expected. Self-report questionnaires are administered on a tablet with mandatory responses for each item, thereby limiting the risk of incomplete questionnaires. Missing data may nevertheless occur in specific situations:

• •for questionnaires or the computerized task, in the event of a technical issue (e.g., software malfunction or device interruption);
• •for biological samples, in cases of insufficient salivary volume or refusal to provide a sample for genetic analysis.

Given these conditions, no imputation or replacement of missing data is performed. The extent and distribution of missing data is documented in a flow chart detailing the data collection flow for each type of variable.

Analysis of the primary outcome: A comparison of attachment-style prevalence (secure vs. insecure) between groups (no-PTSD and PTSD) is applied using a chi-square test of independence.

General design: A descriptive overview of the study population is provided for each group. Categorical variables are summarized using frequencies and percentages. The central tendency and dispersion of continuous variables are reported according to the normality of their distribution, assessed using the Shapiro–Wilk test. When normally distributed, continuous variables are presented as means and standard deviations; otherwise, medians and ranges are reported.Secondary outcomes: For the secondary objectives, Student’s t-tests, Pearson correlations, ANOVA, and mediation analyses (These mediation analyses should be considered hypothesis-generating rather than confirmatory) are used, depending on the outcome measures relevant to each objective. If assumptions for these tests are not satisfied, the appropriate non-parametric alternatives are applied. The biological analyses are exploratory. Statistical power is calculated for each secondary analysis.

Statistical analyses are conducted using Jamovi and easieR (a graphical user interface for R). Correction for multiple comparisons is applied where appropriate using established procedures such as the false discovery rate (FDR; Benjamini–Hochberg). Associations between variables are explored using regression-based approaches, with model specifications guided by the study hypotheses, particularly in secondary analyses. Mediation analyses are conducted using regression-based approaches (e.g., PROCESS macro). Relevant covariates such as age, sex, medication use, and history of traumatic brain injury (TBI) are considered in the analyses. Given the potential non-normal distribution of HRV indices, appropriate transformations (e.g., log-transformation) or non-parametric approaches are used when necessary. Genotype distributions are tested for Hardy–Weinberg equilibrium prior to genetic association analyses.

Protocol validation

Methodological innovation

The AT-HOME protocol was designed to address several methodological challenges related to the study of social cognition in high-stress populations, particularly within military contexts.

First, this protocol was designed to be portable and rapidly deployable in constrained environments, enabling the investigation of social–emotional processes outside traditional laboratory settings. Data collection takes place in operational environments where strict laboratory conditions are not always possible. To address this constraint, the protocol implements standardized procedures to ensure data reliability, including the use of calibrated collection devices, precise recording of sampling times, and controlled pre-analytical storage conditions. This approach allows for the preservation of data quality while achieving high ecological validity. Second, this protocol proposes a reproducible experimental framework integrating, behavioral performance during an emotional recognition task (involving psychological priming, physiological monitoring), psychometric assessments (attachment, emotional contagion, empathy and trauma history) and biological sampling linked to neuropeptides related to social behaviors. This multimodal integration within a single session enables a comprehensive biopsychosocial characterization of emotional processes. Third, our computer based-task was fully pre-tested and iteratively adapted to ensure its feasibility and acceptability among military personnel suffering from post-traumatic stress disorder (PTSD), a population for whom experimental protocols can be particularly difficult to implement. Additionally, we developed a five-minute secure priming procedure specifically designed for use with injured military personnel (high stress population). This priming consists of a professionally recorded audio track, ensuring a high level of standardization and methodological control while remaining appropriate for this population. Fourth, the protocol introduces an experimental integration of attachment-based secure priming within a psychophysiological paradigm. While secure priming has been used in social psychology, it is rarely combined with autonomic nervous system recordings during emotional processing tasks. The present protocol synchronizes priming exposure with continuous EDA and HRV measurements, allowing the investigation of how perceived interpersonal safety modulates emotional processing and physiological regulation. Attachement is also assessed using a combined categorical and dimensional approach, which allows both clinical characterization and quantitative modelling of attachment-related variability. This dual approach improves analytical flexibility for future studies investigating attachment as a moderator of emotional processing. Fifth, the AT-HOME protocol provides detailed synchronization between experimental stimuli presentation, priming procedures and physiological recording, ensuring reproducibility of signal acquisition and facilitating replication of the experimental design by other research teams.

Together, these methodological elements provide a replicable and adaptable framework for investigating social cognition and emotional processing in populations exposed to chronic stress or trauma, in ecological conditions.

Limitations

The protocol presents several limitations, particularly related to data collection in military setting. They include: (1) cross-sectional design, (2) the low representation of female service members, but reflecting the proportion of female personnel in the French Army, (3) use of self-report assessment, (4) the lack of standardization in the timing of saliva sample collection, (5) the need to adapt the order of certain procedures based on participant availability, especially in the PTSD group and (6) generalizability limited to French military.

CRediT authorship contribution statement

Alyson Miquel: Conceptualization, Methodology, Investigation, Writing – original draft. William Chan: Software, Methodology. Christophe Rouquet: Software, Methodology, Investigation. Jules Mison: Software. Margot Jolly: Methodology. Mathieu Coutan: Methodology. Ouamar Ferhani: Methodology, Supervision. Marion Trousselard: Conceptualization, Methodology, Writing – original draft. Anaïs M. Duffaud: Conceptualization, Methodology, Writing – original draft, Supervision, Investigation, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

No data was used for the research described in the article.